Skip to main content



Corrupted blood incident—a not-so-virtual epidemic in a virtual world

On the week before the thanksgiving break we learned about epidemics—the model of the spread of diseases, and estimates of who/how many people would be infected at a given time in a given graph. While we have a lot of plausible models for explaining and estimating what might happen in an epidemic, it is hard for us to do a case study, or a controlled experiment on the development of an epidemic (it would be unethical to create an epidemic for the sake of experiment, and “contained” environments still pose a significant threat to experimenters). In 2005, however, there was an epidemic that affected more than a million people, and researchers got to use it as a model to explain behavior among networks of people during an outbreak.

 

What? Don’t worry, no one died. This happened inside a popular Massive-Multiplayer Online Role-Playing Game (MMORPG), World of Warcraft (WoW for short). 2005 was the heyday of WoW, with over 2 million players worldwide. One day in 2005, a player found a bug related to a particular fight with a monster. This monster could “infect” a player with a deadly disease that would damage the player’s health and would spread to other players in the vicinity. This effect was intended to be temporary and would disappear once the infected character went out of the area that the monster lived, but someone found a way to carry the infection, by infecting his in-game pet animal and sending it back home.

networksimg

 

Here is the contact graph. To explain a little bit more about each nodes, the monster called “Hakkar” is the source of the epidemic. The pets were the first to be infected. The involved characters (pet owners) were infected due to proximity, and by interaction, the non-player characters/players who had some form of immunity became carriers of the disease. This spread to secondary infections to other players, and ended up severely harming low-level players (equivalent of youth/elders/high-risk groups in real life).

 

After this “containment breach”, the virtual epidemic started to become “real.” The disease spread from animals to “human” players, and from players to again animals, and so on. The reproductive number of the disease was over 1, which meant that starting from the first infected person, as the level of the contact tree of that person increased, the number of infected people increased. To fight the infection, players isolated the infected players to less populated areas. Players who had the ability to heal other players tried to cure other players temporarily, and sometimes contacted the disease themselves. Some infected players had “jobs” that they would have to go to, such as in the marketplace and in the bank (yes, the game had its own economy), and usually ended up infecting other players. Scammers sold fake cures to desperate players, and many players infected themselves or other people “just for fun.” This epidemic lasted for a week, when afterwards the game company in question decided to solve the exploit and used “time travel” to turn the world one week back.

 

What happened as a small bug in a game actually proved to be very similar to what happens in real life in the case of a real epidemic. Even though a death in an online game usually is much lighter than a real-life death, people acted exactly similar to how people would actually behave. Even the “dumb” things that people do—treating others even when they know that it is in vain, staying in infected areas just for fun and becoming infected themselves—matched exactly with reported activities during a real-life epidemic.

Citations:

Corrupted Blood: How Epidemiologists Use World of Warcraft to Save Lives. http://all-that-is-interesting.com/corrupted-blood

BBC News. http://news.bbc.co.uk/2/hi/technology/4272418.stm

Reuters: Online “Blood plague” offers lessons for pandemics. http://uk.reuters.com/article/us-flu-virtual-idUKTRE53Q4HI20090427

ET Lofgren, NH Fefferman: The untapped potential of virtual game worlds to shed light on real world epidemics. http://www.sciencedirect.com/science/article/pii/S1473309907702128

Comments

Leave a Reply

Blogging Calendar

November 2016
M T W T F S S
 123456
78910111213
14151617181920
21222324252627
282930  

Archives